Modelling real-time applications based on resource reservation

Santander, 25/09/2010

1

Laura Barros, César Cuevas, Patricia López Martínez,

José M. Drake and Michael González Harbour

Grupo de Computadores y Tiempo Real

Universidad de Cantabria, Spain

{barrosl, cuevasce, lopezpa, drakej, mgh @unican.es}

2ndInternational Workshop on Analysis Tools and

Methodologies for Embedded and Real-time Systems Porto, July 5th, 2011

Funded by the EU under contracts FP7/NoE/214373, and by the Spanish Government under grant TIN2008-06766-C03-03

Modelling real-time applications based

Introduction



Resource Reservation (RR)



Executing each system

thread or communication

session in a

server



Server:

it has assigned a

fraction of the processor

capacity or the

communication network.



Advantages:

•

System robustness

•

Design simplicity

•

Reusability of software

components



MAST



Open source set of tools

to design and analysis of

RT applications



MAST

model

extended

=>

MAST 2



MAST 2

tools

under

development:

•

Possibility of usage of

MAST 1

tools

transforming models

Modelling elements for the resource reservation paradigm I

3 Porto, 5/07/2011 Laura Barros, César Cuevas, Patricia López, J.M. Drake, M.G.Harbour

triggerA

stepA1

stepA2

stepA3

e2efA

tr1

Reactive

model

stepB1

stepB2

triggerB

e2efB

tr2

eA1

eA2

eB1

Modelling elements for the resource reservation paradigm I

triggerA

stepA1

stepA2

stepA3

e2efA

tr1

Reactive

model

stepB1

stepB2

triggerB

e2efB

tr2

processorX

primaryScheduler

mutexU

thread1

Resource

model

thread4

thread3

thread2

eA1

eA2

eB1

Modelling elements for the resource reservation paradigm II

5 Porto, 5/07/2011 Laura Barros, César Cuevas, Patricia López, J.M. Drake, M.G.Harbour

stepB1

stepB2

triggerB

e2efB

tr2

Virtual

platform

triggerA

stepA1

stepA2

stepA3

e2efA

tr1

virtualRsrc

virtualRsrc

virtualRsrc

virtualRsrc

Reactive

model

eA1

eA2

eB1

New classes in resource reservation MAST models II

7 Porto, 5/07/2011 Laura Barros, César Cuevas, Patricia López, J.M. Drake, M.G.Harbour

Attributes of the contract:

VirtualRsrc



Budget

(t

)



Replenishment period

(t

)



Access Deadline

(t

)

Real-time application development

Application

code

Timing

requirements

Application

Description

Application

Workload

« Mast2»

Platform

Description

« Mast2»

RR-model & constr.

(a) 4d_b1+4d_d1≤11.8 4d_b2+4d_d2+2d_a2≤101.8 (b) d_b1=3.1, d_d1=8.8, d_b2=9.54,d_d2=7.2,d_a2=1.2 (c) ...

« Mast2»

RR-model

« Mast1»

RT_Analysis

Model

« Mast1»

RT_Simulation

Model

Analysis

Design

« Mast2»

Exec-model

Execution

Negotiate

virtualPlatform

Vr_Central

(vr1)

Vr_Bus

(vr2)

Vr_Remote

(vr3)

Reactive model of ServoControl example

9 controlTrans: EndToEndFlow T1clockEvent a1 readGoal a2transmitReq a3readSensorPosition a4returnPosition a5 evaluateControl a6transmitControl a7setServoInputs «PeriodicEvent» period=0.01 «Step» SimplOp(wcet=4.0E-4) «Step» message(maxSize=64) «Step» SimplOp(wcet=7.5E-5) «Step» Message(maxSize=256) «Step» SimplOp(wcet=0.8E-3) «Step» Message(maxSize=256) «Step» SimplOp(wcet=5.0E-5) «HardGlobalDeadline» deadline=5.0E-3 refEvent= ”clockEvent” e1 e2 e3 e4 e5 e6 end

Porto, 5/07/2011 Laura Barros, César Cuevas, Patricia López, J.M. Drake, M.G.Harbour

B = a1 + a5

T = T1

B = a2 + a4+a6

T = T1

B = a3 + a7

T = T1

Virtual Platform assignment tool

Period Budget Deadline

Vr_Central(vr1) T1 a1+ a5 ? Vr_Bus(vr2) T1 a2+a4+a6 ? Vr_Remote(vr3) T1 a3+a7 ? Virtual Platform

2

vr1

.t

3

vr2

.t

2

vr3

.t

≤ 9.182 ms

vr1

.t

-(

vr1

.t

-a1)+

vr1

.t

-(

vr1

.t

-a5)+

+

vr2

.t

-(

vr2

.t

-a2)+

vr2

.t

-(

vr2

.t

-a4)+

vr2

.t

-(

vr2

.t

-a6)+

+

vr3

.t

-(

vr3

.t

-a3)+

vr3

.t

-(

vr3

.t

-a7) <

t

GD

wrt (a_x) < vr.t_D– (vr.t_B– a_x)

It is possible to calculate the

wrt

(t

) of

Virtual Platform adjustment

Virtual Resources generation

Iterative search of a set of deadline values that:



makes the application schedulable



verifies the specified restrictions

MAST

Schedulability analysis

Tool

Negotiation tool

Negotiation tool

Vr/SchedRsrc Period Priority Budget Deadline

vr1/centrThr T1 14 1.2 ms 2.51 ms

Vr2/commChannel T1 146 576 bits 0.83 ms

Vr3/remoteThr T1 22 0.125 ms 0.5 ms

Laura Barros, César Cuevas, Patricia López, J.M. Drake, M.G.Harbour Porto, 5/07/2011

Model transformation

Results of analysis using

MAST tools

Complex application: No restrictions. Virtual schedulability analysis

vrx:VirtualSchedulableResource budget=t_B replenismentPeriod=t_R deadline=t_D MAST 2 vrxProc:RegularProcessor (Default attributes) vrxSch:PrimaryScheduler Host=vrxProc Policy=FixedPriorityPolicy vrx:Thread vrxThLoad:Thread vrxProc:EndToEndFlow trg:PeriodicEvent Period=t_D act:Step Thread=vrxLoad Operation= vrxOp vrxOp:SimpleOperation Wcet= t_D-t_B Scheduler=vrxSchParam=Fixed PriorityParam(priority=1) Scheduler=vrxSch Param=FixedPriorityParam( priority=2) MAST 2

ATL

13

ATL

MAST

Schedulability analysis

Tool

Design

Analysis

Model Transformation for the MAST 1 compatibility

MAST 1

MAST 2

Schedulability

VirtualTool

Laura Barros, César Cuevas, Patricia López, J.M. Drake, M.G.Harbour Porto, 5/07/2011

Conclusions



Modelling elements based on RR paradigm with MAST:

•

Scenario:

–

The application can be analysed independently of the current workload

–

The programmers do not require any knowledge about the underlying

platform

•

Solution:

–

MAST tools cover the different phases in the development and execution of

applications based on resource reservation

–

Relying in the availability of a resource reservation middleware installed in

the platform

•

Currently working on:

–

Updating the MAST tools in order to support the new advanced paradigms

for real time systems covered by MAST2

–

Other :



Implementation of the virtual platform assignment tool